XPS8940, Nvidia/BIOS Lock up FINALLY captured....

I have been following these 8940 lock up threads for over a year. I want to say well done on finally finding a culprit. IMO, Dell should have sent everyone having the lock up problem an 8950.

Well, first of all, Support should have been able to do better. Didn't need a complete system replacement, an AMD similar video card would have been a 'cheaper' swap. They refused... and all the way to L3 no less. Only can replace what you bought.

I had problems with the Nvidia card from day 1 18 months ago. PC would boot, the Windows 'flashlight' screen would change to the desktop and within seconds the screen would go black and then about 10 seconds later, the desktop 'bloomed' back and until the next boot I had no problems.

That resulted is in a lot of work for me, clean boots, re-installs, BIOS and Nvidia driver changes. Didn't fix it. 

Finally I ran DDU to clean out the Nvidia drivers while using the Intel HD750, installed Nvidia drivers and the problem went away.

Then last Jan. the BIOS V2.4.0 started the lockups... and continues until this day even. Support is/was useless. The standard stuff, clean boot, re-install, change/try different drivers... it continues to happen.

I was offered a complete system swap, but the terms were I had to basically make no changes to my system... Nope, not handing it over unless I can do a complete CLEAN of the PC and remove ALL licensed s/w and registered ones so I can put them on the replacement. Dell would not allow me to do that, no exchange.

I'm surprised that this is a widespread problem though, that is the error, and when you do Google and look at the pages, many will have 'lockups' and possible symptoms of this. However, a majority it seemed had BSOD's. That or sudden reboots.

The Registry keys mentioned I do not have, and some sites tell you you need to create them. Reluctant to try that though.

I'm hoping others can run the program and see if they get the same or similar data.

I'm well out of warranty and I don't think Dell would help ONE BIT now...

If THIS is the cause, and it is so prevalent on the Nvidia cards, why hasn't Nvidia fixed this? I am on the latest driver that was installed the morning of the lockup no less... Beyond me...

@kras1, what amazes me more, is this link by Adobe of all people:

https://substance3d.adobe.com/documentation/spdoc/gpu-drivers-crash-with-long-computations-tdr-crash-128745489.html 

From July of 2021 no less... from it:

=================

On Windows, this window will appear if Substance 3D Painter detects that the current TDR value is below a specific limit (10 seconds).

Why does the GPU driver crash ?

In order to prevent any rendering or GPU computation from locking up the system, the Windows operating system kills the GPU driver whenever a rendering takes more than a few seconds. When the driver is killed, the application using it crashes automatically. It is not possible to know how long a rendering task or a computation may take (it depends on the GPU, the drivers, the OS, the mesh size, the texture size, etc.), therefore it is not possible to put a limit on how much the computer should process and avoid the crash from the application level.

==================

That led me looking into TDR more (I did not BOLDFACE anything above, it is as posted.

Looking down that page at #5 is the 'supposed' Registry changes needed, I do NOT have those keys.

So I assume then that the driver defaults are used?

If Adobe knew about this, why not DELL (if this is the fix)?

I have a feeling that Dell knows but they can't fix the lock ups without affecting the PC's that do not experience the lock ups. It is weird that not all 8940's with Nvidia cards have this issue. There has to be another common link between the PC's in question. If Adobe knows about this problem then certainly Nvidia and Microsoft do too. It is quite astonishing that no one cares enough to come up with an update to fix it. Not everyone is capable or comfortable with changing the registry on their PC. Either way, Dell should have stepped up and made this right a long time ago. Telling people that they have to leave their PC like it is and send it off to a stranger is ridiculous.

---

@kras1 wrote:

I have a feeling that Dell knows but they can't fix the lock ups without affecting the PC's that do not experience the lock ups. It is weird that not all 8940's with Nvidia cards have this issue. There has to be another common link between the PC's in question.

---

Well, if that Adobe web site IS correct, that is the cause, then the problem is  within NVidia. See, the TDR setting it seems is within the Registry. That whole section was probably written by the Nvidia Device Driver install. So either with no key to set the TDR is uses the default. However, IF that is the cause and the fix it to set it correctly, NVidia and Dell are to blame. Again, this is my guess as to WHY, the CAUSE, well WHOCRASHED indicates who's file was at fault.

We have here recently in a thread where the user had TWO XPS's, one from mid-2022 and the other just recently purchased. Both have NVidia card but different models. 5 Monitors were used and the user upgraded to the new one with a more powerful NVidia card and Intel i7 to an i9 I recall. The new unit locks up.

Probing, it turns out the older unit is a 10th Gen. Intel, the new one, 11thGen... and that meant different motherboards and Northbridge and Southbridge chipsets. Probably the chipsets are part of the problem.

So there could be many XPS8940's with 10th Gen Intel out there, for a few months Dell did ship them configured that way. Suspect those are the ones that do not fail.

It isn't Windows either as Linux users have had lockups as well. Purely a h/w situation, and specific to NVidia cards only, but caused initially by changed code in BIOS V2.4.0 and later versions.

I'm not that willing to change something in the registry I'm not sure of why it should be changed, or to what? I'm hoping a Dell person looks at this, thinks it could be at least something to be investigated and works on an answer for this.

Time will tell, that or someone (it might be me at some point) makes the 2 Registry additions and after a few weeks or months declares it to be the FIX!!!!

Hi Irv,

Great job troubleshooting.   As I noted before, I have not had these lockups on my XPS8940 with its Nvidia RTX 2060 Super.    A search of my Windows directory shows nvlddmkm.sys found in:

C:\Windows\System32\DriverStore\FileRepository\nvdd.inf_amd64_167c7688cc7d203a\nvlddmkm.sys w/ date last modified of 11/1/2022 8:34AM, size: 42,783KB

@ishmael,

I keep up with the NVidia releases for the Game Ready driver.

So mine is newer, and might be a difference if you are using the Studio Driver:

C:\>dir nvlddmkm.sys /s
Volume in drive C is OS
Volume Serial Number is 9A97-C0F4

Directory of C:\Windows\System32\DriverStore\FileRepository\nvddi.inf_amd64_52d9834be3be904c

01/18/2023 12:48 AM 58,078,744 nvlddmkm.sys

It is 15KB or so larger.

What CPU do you have? Just curious?

Also, in REGEDIT, if you go here,

Computer\HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers

Does it look like this?

Or are you missing the last 2?

@RoHe 

Ron, you SURE DID, and every once in a while I did run it. NEVER before did I get anything found.

This was the first time. I've had it on the XPS from possibly the first time you suggested it. This time it did need and update. Maybe it changed where it looked for more problems?

I suspect this error is the equivalent of the 'falling off the bus' errors Linux users get?

@ispalten -  IIRC, I suggested running WhoCrashed eons ago in one of the many XPS 8940 freeze threads...

@ispalten - Maybe this is just a different type of freeze, since WhoCrashed found something this time.

Now you have to figure out what to do about this one.

if you updated the NVidia driver after the previous crash, maybe revert to the prior version. And run WhoCrashed if PC crashes again after you revert.

If WhoCrashed doesn't find anything with the previous driver, you may have to conclude the newer one introduced some new issue...

@RoHe I've been having lock ups regularly on all drivers. A few driver installs before the last two even would lock up as the install completed.

Since in the beginning WHOCRASHED never found anything I didn't bother running it recently.

Since it happened so close to the last install I started doing some Web searches about Nvidia problems and and found some about DTR delay and they could cause lock ups.

For whatever reason, since the lock up happened so close after installing the latest driver I ran WC. Much to my surprise, after updating WC, i got the results I posted. Time matched exactly to my lock up and reboot times.

Surprisingly, if one Googles what WC suggested, well, this IS quite common and NOT only on Dell equipment either. Almost all have no real guaranteed 100% cure either, and some, like I said, have BSOD's vs. a lock up.

DTR setting have also been implicated in some AMD problems as well.

What surprises me is not Registry setting entries for DTR that it seems some either have or found to need them so they create them? One would think if these were needed, the Device Driver would make the entries?

Now it is possible the driver has its own data, and based on what it sees when it installs, it determines what it thinks is the proper time and saves that? I'm not having much luck finding why those are not in the Registry already?

I'm not sure how I could tell it would fix it, other then making those entries, and the problem there is WHAT is the value? Too small, and assume this would correct it, I'd fail. Not sure what the consequences are of making too large either? Performance hit? Lagging too much?

Hey, Google this if you want to get confused, "nvidia what should tdr setting be?", too many and some have different opinions.

Here, read this one, https://docs.nvidia.com/gameworks/content/developertools/desktop/timeout_detection_recovery.htm and from it:

-------------------------------

TDR stands for Timeout Detection and Recovery. This is a feature of the Windows operating system which detects response problems from a graphics card, and recovers to a functional desktop by resetting the card. If the operating system does not receive a response from a graphics card within a certain amount of time (default is 2 seconds), the operating system resets the graphics card.

Before TDR existed, problems of this nature would have resulted in a system freeze and required a reboot of the operating system. If TDR is enabled and you see the TDR error message, "Display driver stopped responding and has recovered," this means that the Windows operating system reset the display driver.

-------------------------------

Nvidia page. so since I don't have that set in the Registry, does that mean "Before TDR existed, problems of this nature would have resulted in a system freeze and required a reboot of the operating system." as I don't have it set the same as not having it?

Of course, this MS link, https://learn.microsoft.com/en-us/windows-hardware/drivers/display/timeout-detection-and-recovery, does say it will freeze, but it is only a few seconds and come back, implying the GPU is working on intense Graphic (which usually is NOT the case).

Irv:   Here's my CPU/GPU info:

CPU: 11th Gen Intel(R) Core(TM) i7-11700 @ 2.50GHz, 2496 Mhz, 8 Core(s), 16 Logical Processor(s)

GPU: Nvidia RTX 2060 Super, 8GB / Driver Version: 517.66 (Gaming Version)

And yes, it appears I am missing those last two entries in the registry

@ishmael thanks for the info.

Where did you get that driver from? That I suspect is the Dell OEM one from Dec.??? https://www.dell.com/support/home/en-us/drivers/driversdetails?driverid=nphhm&oscode=w2021&productcode=xps-8940-desktop  I guess I'll give it a try.

Seems older than my Driver Version, 31.0.15.2824 which is 528.24 from 1/24/2023.

Well, I did install the same one from Dell, did the 'clean install'.

So doing so I get a H/W Error!!! Only my mouse could be moved after I closed the Install that said it completed.

Can't find much about it either:

Rebooted and so far all seems OK. We'll see if I get a lock up, but can take a long time possibly.

---

@RoHe wrote:

@ispalten -  IIRC, I suggested running WhoCrashed eons ago in one of the many XPS 8940 freeze threads...

---

Ron,

I was contacted via PM from Dell to USE the 'official' Dell driver. It was 31.0.15.1766. Support can't run non-Dell programs it seems, so I was to change drivers (Clean Install was done) and then test. Well, it didn't take too long and I locked up under that driver.

So I ran WhoCrashed, nope, no data other than the OLD one from last Sat. No NEW data was written.

So, either only under certain circumstances it seems will data be written or it is a function of the specific Driver being used? Like the most recent from NVidia maybe?

That driver problem is all over the web for a variety of in use h/w, not only Dell specific h/w or drivers. Possible Nvidia is putting more 'debug' into the drivers to capture why?

Today I was also notified by Dell Update of a new Nvidia driver released in Dec. The one Dell asked me to install and I did was from Oct. Looks like the new one is 31.0.15.2698. Meanwhile NVidia most recent is 31.0.15.2824 which I was running when I got the dump info for WhoCrashed.

I don't understand why there could be no data?

Sort of 'implies' there is more than one scenario that the XPS locks up under?

All this was done under BIOS V2.10.0. Typical problem on that BIOS... I can get weeks and I can get days or even more than once a day lock ups with it. Totally random with it happening doing different things, sigh.